32 lines
844 B
Python
32 lines
844 B
Python
import joblib
|
|
import numpy as np
|
|
|
|
# Load model
|
|
model = joblib.load('ml_assets/ml_pipeline_internal.pkl')
|
|
|
|
# Ekstrak komponen
|
|
vec = model.named_steps['tfidf']
|
|
clf = model.named_steps['clf']
|
|
|
|
# Dapatkan vocabulary
|
|
feature_names = vec.get_feature_names_out()
|
|
|
|
# Dapatkan bobot koefisien
|
|
coefficients = clf.coef_
|
|
intercepts = clf.intercept_
|
|
|
|
# Tampilkan bobot untuk kata-kata di korpus mini
|
|
target_words = ['web', 'develop', 'analyst']
|
|
|
|
print("BOBOT KOEFISIEN RIIL:")
|
|
print("=" * 60)
|
|
for word in target_words:
|
|
if word in feature_names:
|
|
idx = list(feature_names).index(word)
|
|
print(f"\nKata: '{word}'")
|
|
for i, cls in enumerate(clf.classes_):
|
|
print(f" Kelas {cls}: {coefficients[i][idx]:.4f}")
|
|
|
|
print("\nINTERSEP (Bias):")
|
|
for i, cls in enumerate(clf.classes_):
|
|
print(f" Kelas {cls}: {intercepts[i]:.4f}") |